One of the key criteria in selecting a climate scenario was that it should be physically plausible. This criterion also applies to the relationship between climate and non-climatic scenarios. Thus, projections of climate should be consistent both with projections of atmospheric composition and the emissions scenarios upon which they are based, as well as with "downstream" projections of sea-level rise. Carbon dioxide concentration is one of the most important of the non-climatic factors to consider. Besides being a major greenhouse gas that influences the climate, it is also of great importance for plant growth and productivity. Thus, it is important that appropriate levels of CO2 concentration are used in conjunction with a given climate change. Unfortunately, this has been a source of some confusion in past impact studies, especially with regard to the timing of CO2 doubling. The following points are worth noting:

Equilibrium GCM 2 x CO2 experiments commonly assume a radiative forcing equivalent to a doubling of CO2 concentration (for example from 300 ppmv to 600 ppmv). In fact the absolute concentrations are not especially important, as the temperature response to increasing CO2 concentration is logarithmic - a doubling from 500 to 1000 ppmv would have approximately the same climatic effect.
Equilibrium 2 x CO2 experiments are usually interpreted as representing an equivalent 2 x CO2 atmosphere, in which the combined effects of CO2 and other greenhouse gases on the earth's radiation balance are equivalent to the effect of doubling CO2 alone. This equivalent doubling is expected to occur several decades before an actual CO2 doubling.
The climate has a lag time of several decades in its response to an equivalent CO2 doubling, as represented in transient AOGCM experiments. Thus, at the time of equivalent doubling, the climate will not have realized its full, equilibrium response to the forcing.
In CO2 enrichment experiments with plants, a doubled CO2 environment relative to ambient is typically imposed (levels of 660 ppmv or 700 ppmv are popularly adopted). In contrast to GCMs, the absolute CO2 concentration is of importance to plants, because their response to CO2 is commonly non-linear.

Similarly, relative sea-level rise is predominantly a result of global warming, so this should also be consistent with the estimates of climate change.

Scenario reporting

In this section suggestions are put forward concerning the presentation and reporting of impact assessments, especially concerning the use of scenarios. Adherence to some of these basic guidelines will greatly assist the reviewing and synthesis of impact studies.

Appropriate citation of sources

Out of courtesy to the scientists involved, the original sources of the baseline data and scenarios used should be cited correctly. For example, although the Data Distribution Centre will be providing data from AOGCMs, the correct sources to cite in referring to these models are publications by the modelling groups themselves, not the DDC or these Guidelines. The DDC has documented each of the models, so the relevant information is readily available. Similarly, the sources of non-climatic scenarios should also be referenced correctly . If components of these scenarios are to be applied (for example, regional population projections) then the original source of the projections should be cited . Again, the DDC provides guidance on these.

Use of standard notation

Special care should be taken to adopt conventional notation when referring to individual GCM experiments. There are many versions of the the same or similar models in circulation, so it is important to identify models using an accepted acronym. Again, the DDC provides guidance on these.

Description of methods

The methods adopted to select, interpret and apply the scenarios should be described in full, with proper citation to comparable previous studies employing similar methods. This information is important for evaluating and comparing different impact studies.

Presentation of results

Impact studies that employ scenarios should indicate, where possible, the statistical significance of the results. For example, regional scenarios of climate change should be compared with natural variability in the baseline observations or control simulation. Similarly, the impacts of these scenarios should be contrasted with the impacts of natural variability.

Consideration of uncertainties

At each stage of an impact assessment, there should be a full and proper dicussion of the key uncertainties in the results, including those attributable to the input data, impact models, climate scenarios and non-climatic scenarios. A rigorous sensitivity analysis can be very helpful in identifying some of the major uncertainties. It is also recommended that users should design and apply multiple scenarios in impact assessments, where these multiple scenarios span a range of possible future climates, rather than designing and applying a single "best-guess" scenario